Stable Video Diffusion XT is useful for trying a little movement in an existing still: a short camera drift or an atmospheric shot. It generates 25 frames at 576 × 1024, giving local experiments a clearly bounded target. Text prompts cannot direct the released model, movement may be minimal or absent, and faces can break down—poor foundations for character close-ups. Keep it as a comparison checkpoint or try it for brief non-explicit inserts, inspecting every frame before use. Dialogue, planned sequences and precise animation need other tools, and commercial use depends on the checkpoint’s license.
Commercial use & license
The checkpoint uses Stability AI’s Community License. Commercial use is subject to registration and revenue conditions, and all use must follow its acceptable-use policy.
Adult content rules
Stability’s September 30, 2026 acceptable-use policy prohibits content relating to sexual intercourse, sexual acts or sexual violence. It applies to self-hosted code and weights under Research, Community or Enterprise licenses, as well as use through third-party hosts.
What you need
A 576 × 1024 input still, the XT checkpoint and a compatible image-to-video pipeline. XT generates 25 frames at that resolution. GPU memory needs vary with the implementation and memory optimizations.
Ranks are recommendations for the task described, based on reviewed documentation. Model ranks cover non-explicit development work. We have not run practical benchmarks. Check content policies and commercial terms separately. How the ranks work →
Sources & further reading +
Last checked 5 October 2026.